Figure 2 | Vectorial DNN for polarized object perception. (IMAGE)
Caption
Figure 2 | Vectorial DNN for polarized object perception. a, Schematic of Poincaré-sphere-perceptive classification, where nine categories of images in three polarization states (red arrows and dots) are allocated to three detection areas with a fixed horizontal analyser. b, Experimental results of polarization-multiplexed image classification in the second polarization channel. c, Experimental confusion matrices of the three multiplexed image classification tasks. d, Schematic of polarized object perception, where objects with different polarization features can be sensed and classified by the vectorial DNN. e, Experimental results of polarized object perception for the second category. f, Experimental confusion matrix of the polarized object perception task. g, Envisioned application of the vectorial DNN in real-time biological sample analysis and pathological diagnosis. h, Optical setup of a convolutional vectorial DNN for real-world cell classification. P1-P2, polarizers; L1-L3, lenses; LC1-LC2, photo-rewritable LC devices. i, Experimental results of sensing and classifying two types of plant cells. The scale bars in the micrographs represent 400 μm.
Credit
Peng Chen et al.
Usage Restrictions
Credit must be given to the creator.
License
CC BY